Multi-Exposure Image Fusion: Perceptual Quality Assessment and Dataset Overview

This paper focuses on perceptual quality assessment for multi-exposure image fusion techniques. Two datasets are used in this research:

1. 'ICVL Multi-exposure Fusion Dataset': This dataset contains 11 pairs of multi-exposure images, with each pair including three images captured under different exposures. These images represent diverse scenes, encompassing both indoor and outdoor environments. Each pair has a reference image, enabling the evaluation of fusion algorithms' performance.

2. 'MIT-Adobe FiveK Dataset': This dataset comprises 5000 high-resolution images sourced from various cameras. These images encompass a wide range of scenes, including indoor, outdoor, and natural landscapes. This dataset is employed to assess the visual impact and image quality of fusion algorithms.